Prognostic Impact of Hyperpolypharmacy Due to Noncardiovascular Medications in Patients After Acute Decompensated Heart Failure　― Insights From the Clue of Risk Stratification in the Elderly Patients With Heart Failure (CURE-HF) Registry ―

Abstract

Background: Hyperpolypharmacy is associated with adverse outcomes in older adults, but because literature on its association with cardiovascular (CV) outcomes after acute decompensated heart failure (ADHF) is sparse, we investigated the relationships among hyperpolypharmacy, medication class, and death in patients with HF.

Methods and Results: We evaluated the total number of medications prescribed to 884 patients at discharge following ADHF. Patients were categorized into nonpolypharmacy (<5 medications), polypharmacy (5–9 medications), and hyperpolypharmacy (≥10 medications) groups. We examined the relationship of polypharmacy status with the 2-year mortality rate. The proportion of patients taking ≥5 medications was 91.3% (polypharmacy, 55.3%; hyperpolypharmacy, 36.0%). Patients in the hyperpolypharmacy group showed worse outcomes than patients in the other 2 groups (P=0.002). After multivariable adjustment, the total number of medications was significantly associated with an increased risk of death (hazard ratio [HR] per additional increase in the number of medications, 1.05; 95% confidence interval [CI], 1.01–1.10; P=0.027). Although the number of non-CV medications was significantly associated with death (HR, 1.07; 95% CI, 1.02–1.13; P=0.01), the number of CV medications was not (HR, 1.01; 95% CI, 0.92–1.10; P=0.95).

Conclusions: Hyperpolypharmacy due to non-CV medications was associated with an elevated risk of death in patients after ADHF, suggesting the importance of a regular review of the prescribed drugs including non-CV medications.

Heart failure (HF) is a growing healthcare problem in the aging population and is associated with a significant risk for recurrent cardiovascular (CV) events. Comorbidities are common in patients with HF and may contribute to an increased risk of adverse outcomes.^1–4 Medications for comorbidity management are prescribed in accordance with current guidelines, often leading to an increase in the number of medications consumed. Polypharmacy has become increasingly relevant for patients with HF, due to the expanding armamentarium of guideline-directed medical therapies for treating HF.^5–7 Polypharmacy is the common use of ≥5 medications daily; however, taking ≥10 medications is notable, representing a condition described as hyperpolypharmacy in the geriatric and pharmacology literature.^8–10 Polypharmacy is reported to be associated with adverse clinical outcomes in older adults,^11,12 with a reported prevalence of 55% among 558 HF hospitalizations in patients aged ≥65 years in the USA,⁵ and 54% in a large cohort of patients with HF and a preserved ejection fraction who were enrolled in the “Treatment of Preserved Cardio Function Heart Failure With an Aldosterone Antagonist” (TOPCAT) trial.¹³ In the TOPCAT trial, compared with nonhyperpolypharmacy, hyperpolypharmacy was associated with increased risks of hospitalization for any reason and serious adverse outcomes; however, there was no significant association between polypharmacy status and death.¹³ In the Kyoto Congestive Heart Failure (KCHF) registry, a Japanese observational study, the prevalence of hyperpolypharmacy among 2,578 patients with acute decompensated HF (ADHF) who were ambulatory at hospital discharge was 37.5%; hyperpolypharmacy was associated with greater risks of death and rehospitalizations within a 1-year period.¹⁴ The clinical relevance of hyperpolypharmacy has been proposed, but data on its prognostic effect on death in HF remain limited. Furthermore, there is a paucity of data on the associations between the number of medications stratified by medication class (CV medications vs. non-CV medications) and adverse events.

The objectives of this study were to assess the prevalence of hyperpolypharmacy and to clarify the association between polypharmacy status stratified by medication class (CV medications vs. non-CV medications) and death after ADHF.

Methods

Patients and Study Design

This was a post-hoc analysis of data from the Clue of Risk Stratification in the Elderly Patients With Heart Failure (CURE-HF) registry, which was a prospective, multicenter, cohort study conducted in the Nagano Prefecture, Japan.^15–17 It enrolled 1,036 consecutive patients who were hospitalized with a primary diagnosis of ADHF and discharged after treatment at 13 institutions between July 2014 and August 2019. The exclusion criteria were patients aged <20 years and those with acute coronary syndromes. After admission, medical therapy was initiated at the investigator’s discretion at each local site. Baseline clinical data, including demographic characteristics, past medical history, laboratory data, and echocardiography findings, were assessed during the compensated state of HF. All-cause death was tracked for 2 years. Follow-up data were obtained either from hospital charts, through direct contact with the patients or from the referring physicians. To ensure an accurate assessment of clinical events, additional information was obtained from visits or telephonic conversations with living patients or their family members and from medical records obtained from other hospitals, as necessary, between June and August 2021. These events were fully anonymized before analysis by the investigators, who were blinded to the participants. The study was approved by each participating institutional review board or ethics committee. All study participants provided written informed consent prior to enrollment. This study was performed in accordance with the Declaration of Helsinki and registered in the University Hospital Medical Information Network (UMIN 000024470).

Of the 1,036 patients with ADHF, we excluded 142 patients with missing clinical data (albumin, B-type natriuretic peptide [BNP], creatinine, hemoglobin, left ventricular ejection fraction [LVEF], and events) and 10 patients with missing follow-up data. The final study group comprised 884 patients (85.3%) (Figure 1). We evaluated the total number of medications prescribed at discharge. Based on prior studies,^8,9,13 polypharmacy status was categorized into 3 groups: nonpolypharmacy (0–4 medications, n=77), polypharmacy (5–9 medications n=489), and hyperpolypharmacy (≥10 medications, n=318), as captured on the case report forms. We defined medications for hypertension, dyslipidemia, HF, coronary artery disease, stroke, peripheral artery disease, and atrial fibrillation as CV medications. For instance, renin-angiotensin-aldosterone system inhibitors (angiotensin-converting enzyme inhibitors (ACEIs), angiotensin-II receptor blockers (ARBs), and mineralocorticoid receptor antagonists (MRAs)), β-blockers, diuretics (loop, thiazide, and tolvaptan), calcium-channel blockers, digoxin, antiarrhythmics, pimobendane, nitrates, nicorandil, statins, and oral anticoagulants/antiplatelets were defined as CV medications.¹³ In recent years, sodium glucose cotransporter (SGLT)-2 inhibitors have been used as a standard medication for HF,^6,7,18–20 so we defined them as CV medications in this study, although they were used as hypoglycemic drugs during the registration period of this study. The following CV medications were defined as HF medications: ACEIs, ARBs, MRAs, β-blockers, diuretics (loop, thiazide, and tolvaptan), digoxin, amiodarone, nitrates, nicorandil, and SGLT-2 inhibitors.⁴ We did not collect data on the medications administered via eye drops, suppositories, ointments, and plasters. We counted combined products (polypill) as each single-ingredient product separately. Over-the-counter medications, vitamins, and mineral supplements administered on a daily basis were not included in this study.

Figure 1.

Flowchart of analysis from the Clue of Risk Stratification in the Elderly Patients With Heart Failure (CURE-HF) registry. BNP, B-type natriuretic peptide; Cre, creatinine; CV, cardiovascular; Hb, hemoglobin; HF, heart failure; LVEF, left ventricular ejection fraction.

We assessed the following comorbidities using the Charlson comorbidity index and the Elixhauser comorbidity index:^21,22 hypertension, dyslipidemia, diabetes mellitus, hyperuricemia, atrial fibrillation, history of coronary artery disease, cerebrovascular disease, peripheral artery disease, chronic kidney disease (estimated glomerular filtration rate [eGFR] <60 mL/min/1.73 m²), anemia (hemoglobin <13 g/dL for men and <12 g/dL for women, based on the World Health Organization [WHO] criteria), obesity (body mass index [BMI] ≥30 kg/m²), chronic obstructive pulmonary disease, dementia, and malignant tumors.

Statistical Analysis

Continuous variables are summarized as mean±standard deviation or as median with interquartile range. Categorical variables are expressed as frequency and percentage. Clinical characteristics across the polypharmacy categories were compared using linear regression for continuous normally distributed variables, Jonckheere-Terpstra non-parametric trend tests for continuous nonnormally distributed variables, and Pearson’s χ² test for categorical variables. Clinical data were compared between the nonhyperpolypharmacy and hyperpolypharmacy categories using an unpaired Student’s t-test or a Mann-Whitney U-test. The following clinical factors associated with hyperpolypharmacy were assessed using backward stepwise regression models (P value threshold=0.05): age, sex, LVEF, New York Heart Association (NYHA) functional class at discharge, prior hospitalization for HF, hemoglobin, albumin, eGFR, BNP, and the total number of comorbidities (range, 0–14; comorbidities included hypertension, dyslipidemia, diabetes mellitus, hyperuricemia, atrial fibrillation, history of coronary artery disease, history of cerebrovascular disease, history of peripheral artery disease, chronic kidney disease [eGFR <60 mL/min/1.73 m²], anemia [hemoglobin <13 g/dL for men and <12 g/dL for women, based on the WHO criteria], obesity [BMI ≥30 kg/m²], chronic obstructive pulmonary disease, dementia, and malignant tumors). Kaplan-Meier curves were calculated from baseline (date of discharge) to the incidence of all-cause death and compared using a log-rank test. Poisson models were used to estimate the incidence rates. A multivariable Cox proportional hazards regression analysis was performed to estimate the association between polypharmacy categories and subsequent clinical events; the following 21 covariates were included: age, sex, LVEF, NYHA functional class at discharge, prior hospitalization for HF, hypertension, dyslipidemia, diabetes mellitus, hyperuricemia, atrial fibrillation, history of coronary artery disease, history of cerebrovascular disease, history of peripheral artery disease, obesity (BMI ≥30 kg/m²), chronic obstructive pulmonary disease, dementia, malignant tumor, hemoglobin, albumin, eGFR, and BNP. Patients were most frequently categorized into the polypharmacy group (5–9 medications), thus, the polypharmacy group was defined as the reference group for the Cox analyses, and estimated hazard ratios (HRs) and 95% confidence intervals (CIs) were calculated. Because the definitions for polypharmacy varied, we also performed a sensitivity analysis wherein patients were classified into the hyperpolypharmacy group (≥10 medications) and nonhyperpolypharmacy group (<10 medications). To assess the relationship of class-specific medications with death, we divided patients into 2 groups (CV and non-CV medications) in the Kaplan-Meier curves according to the median number of these medications prescribed, and conducted multivariable Cox analyses; we computed the HRs per 1 medication increase with 95% CIs. We also conducted subgroup analyses to evaluate the effect of hyperpolypharmacy on the incidence of all-cause death in subpopulations based on age (<80 vs. ≥80 years), sex, NYHA functional class (I–II vs. III–IV), HF etiology (ischemic vs. nonischemic), hypertension, diabetes mellitus, atrial fibrillation, and LVEF (<40% vs. ≥40%). All statistical analyses were performed using SPSS version 28.0 (SPSS; Chicago, IL, USA) and STATA version 14.1 (Stata Corp; College Station, TX, USA).

Results

The baseline clinical characteristics of the 884 patients included in this study are listed in Table 1. The median patient age was 81 years, and 55.1% of the patients were male. The demographic findings did not differ significantly between the 884 patients included in this study and the 152 patients excluded from this study (Supplementary Table 1). The distributions of the total, CV, and non-CV medications at baseline are shown in Figure 2. The median [25th, 75th percentiles] total, CV, non-CV, HF, and non-HF medications were 8 [6, 10], 5 [4, 6], 3 [2, 5], 3 [3, 4], and 5 [3, 7], respectively. The nonpolypharmacy, polypharmacy, and hyperpolypharmacy groups comprised 77 (8.7%), 489 (55.3%), and 318 (36.0%) patients, respectively (Figure 1, Table 1). The proportion of patients taking ≥5 medications was 91.3% (polypharmacy, 55.3%; hyperpolypharmacy, 36.0%). Patients in the hyperpolypharmacy group were more likely to have a worse NYHA functional class and a lower hemoglobin level and eGFR. The total number of comorbidities (range; 0–14) was significantly higher in the hyperpolypharmacy group; common comorbidities included dyslipidemia, diabetes mellitus, hyperuricemia, atrial fibrillation, history of coronary artery disease, history of cerebrovascular disease, and malignant tumors. Furthermore, prior hospitalization for HF was also common in this group. In the multivariable logistic analysis, the number of comorbidities was strongly associated with hyperpolypharmacy (adjusted odds ratio per an additional increase in the number of comorbidities, 1.33; 95% CI, 1.21–1.45; P<0.001; Supplementary Table 2). When examining medications by class, we observed that the number of total medications increased with an increase in the CV and non-CV medications (Figure 3).

Table 1. Baseline Characteristics of Patients According to Total Number of Medications

	Non–polypharmacy (<5 medications) (n=77, 8.7%)	Polypharmacy (5–9 medications) (n=489, 55.3%)	Hyperpolypharmacy (≥10 medications) (n=318, 36.0%)	P value (for trend)
Total no. of medications (range; 0–20)	4 [3–4]	7 [6–8]	12 [10–13]	–
Total no. of CV medications (range; 0–16)	3 [2–4]	5 [4–6]	6 [5–7]	–
Total number of non-CV medications (range; 0–14)	0 [0–1]	2 [1–3]	6 [4–7]	–
Age, years	78 [63–87]	82 [7–87]	81 [74–87]	0.31
≥65, n (%)	56 (72.7)	423 (86.5)	294 (92.5)	<0.001
≥75, n (%)	45 (58.4)	347 (71.0)	230 (72.3)	0.052
≥80, n (%)	38 (49.4)	273 (55.8)	181 (56.9)	0.48
Male sex, n (%)	48 (62.3)	266 (54.4)	173 (54.4)	0.41
BMI	22.0±4.9, n=77	21.5±4.6, n=483	21.7±3.9, n=317	0.26
NYHA functional class (III or IV), n (%)	15 (19.5)	92 (18.8)	79 (24.8)	0.07
Prior heart failure hospitalization, n (%)	10 (13.0)	132 (27.0)	127 (39.9)	<0.001
Etiology: Ischemic cardiomyopathy, n (%)	5 (6.5)	99 (20.2)	120 (37.7)	<0.001
Comorbidities
Total no. (range; 0–14 comorbidities)	2 [2–3]	4 [3–5]	4 [3–6]	<0.001
Hypertension, n (%)	43 (55.8)	329 (67.3)	199 (62.6)	0.096
Dyslipidemia, n (%)	9 (11.7)	97 (19.8)	113 (35.5)	<0.001
Diabetes mellitus, n (%)	9 (11.7)	122 (24.9)	125 (39.3)	<0.001
Hyperuricemia, n (%)	7 (9.1)	84 (17.2)	66 (20.8)	0.049
Atrial fibrillation, n (%)	21 (27.3)	264 (54.0)	161 (50.6)	<0.001
History of coronary artery disease, n (%)	5 (6.5)	82 (16.8)	114 (35.8)	<0.001
History of cerebrovascular disease, n (%)	6 (7.8)	68 (13.9)	63 (19.8)	0.011
History of peripheral artery disease, n (%)	5 (6.5)	45 (9.2)	33 (10.4)	0.56
Chronic kidney disease (eGFR <60 mL/min/1.73 m2), n (%)	45 (58.4)	374 (76.5)	263 (82.7)	<0.001
Anemia (male: <13 g/dL, female: <12 g/dL), n (%)	30 (39.0)	272 (55.6)	213 (67.0)	<0.001
Obesity (BMI ≥30 kg/m2), n (%)	5 (6.5)	20 (4.1)	9 (2.8)	0.30
COPD, n (%)	3 (3.9)	19 (3.9)	21 (6.6)	0.20
Dementia, n (%)	9 (11.7)	65 (13.3)	37 (11.6)	0.76
Malignant tumor, n (%)	5 (6.5)	25 (5.1)	31 (9.7)	0.039
CV medications
ACEIs, n (%)	30 (39.0)	197 (40.3)	124 (39.0)	0.93
ARBs, n (%)	10 (13.0)	148 (30.3)	115 (36.2)	<0.001
ACEIs/ARBs, n (%)	40 (51.9)	341 (69.7)	238 (74.8)	<0.001
β-blockers, n (%)	36 (46.8)	347 (71.0)	251 (79.2)	<0.001
MRAs, n (%)	29 (37.7)	253 (51.7)	196 (61.6)	<0.001
Calcium-channel blockers, n (%)	9 (11.7)	162 (33.1)	110 (34.6)	<0.001
Loop diuretics, n (%)	53 (68.8)	396 (81.0)	295 (92.8)	<0.001
Furosemide, n (%)	33 (42.9)	222 (45.4)	142 (44.7)	0.91
Azosemide, n (%)	16 (20.8)	153 (31.3)	156 (49.1)	<0.001
Thiazide diuretics, n (%)	2 (2.6)	23 (4.7)	37 (11.6)	<0.001
Tolvaptan, n (%)	1 (1.3)	77 (15.7)	104 (32.7)	<0.001
Digitalis, n (%)	0 (0.0)	22 (4.5)	32 (10.1)	<0.001
Amiodarone, n (%)	0 (0.0)	24 (4.9)	39 (12.3)	<0.001
Other antiarrhythmics, n (%)	1 (1.3)	32 (6.5)	35 (11.0)	0.006
Pimobendane, n (%)	1 (1.3)	16 (3.3)	35 (11.0)	<0.001
Nitrates, n (%)	5 (6.5)	24 (4.9)	45 (14.2)	<0.001
Nicorandil, n (%)	1 (1.3)	8 (1.6)	27 (8.5)	<0.001
Statins, n (%)	4 (5.2)	102 (20.9)	130 (40.9)	<0.001
Antiplatelets, n (%)	4 (5.2)	129 (26.4)	166 (52.2)	<0.001
Aspirin, n (%)	3 (3.9)	78 (16.0)	123 (38.7)	<0.001
Warfarin, n (%)	7 (9.1)	127 (26.0)	92 (28.9)	0.002
DOAC, n (%)	15 (19.5)	167 (34.2)	96 (30.2)	0.030
SGLT-2 inhibitors, n (%)	2 (2.6)	19 (3.9)	15 (4.7)	0.67
Non-CV medications
Proton pump inhibitors, n (%)	13 (16.9)	254 (51.9)	226 (71.1)	<0.001
NSAIDs, n (%)	1 (1.3)	7 (1.4)	12 (3.8)	0.077
Oral hypoglycemic agents, n (%)	5 (6.5)	71 (14.5)	93 (29.2)	<0.001
Metformin, n (%)	1 (1.3)	17 (3.5)	7 (2.2)	0.40
Pioglitazone, n (%)	0 (0.0)	5 (1.0)	0 (0.0)	0.13
DPP-IV inhibitors, n (%)	2 (2.6)	49 (10.0)	70 (22.0)	<0.001
Sulfonylurea, n (%)	1 (1.3)	9 (1.8)	17 (5.3)	0.012
Glinide, n (%)	0 (0.0)	2 (0.4)	4 (1.3)	0.27
α-glucosidase inhibitors, n (%)	0 (0.0)	11 (2.2)	18 (5.7)	0.007
Insulin, n (%)	0 (0.0)	9 (1.8)	23 (7.2)	<0.001
Laboratory and echocardiography data
Hemoglobin (g/dL)	13.2±2.5	12.4±2.4	11.8±2.1	<0.001
Albumin (g/dL)	3.5±0.5	3.4±0.5	3.5±0.5	1.00
Creatinine (mg/dL)	1.0±0.7	1.3±0.8	1.4±0.8	<0.001
eGFR (mL/min/1.73 m2 surface area)	57.3 [43.0, 70.7]	46.5 [34.9, 59.1]	40.0 [30.0, 54.0]	<0.001
Sodium (mEq/L)	140±4	139±4	139±4	0.15
Potassium (mEq/L)	4.4±0.6	4.4±0.5	4.3±0.5	0.31
Uric acid (mg/dL)	7.2±1.9	7.2±2.1	6.8±2.0	0.011
BNP (pg/mL)	239 [129, 489]	286 [138, 517]	306 [143, 533]	0.28
HbA1c (%)	6.0±0.6	6.1±0.8	6.2±0.9	0.033
LDL-C (mg/dL)	109±30	98±31	91±30	<0.001
LVEF (%)	46.8±16.6	49.0±15.8	48.2±16.2	0.94
HFpEF (LVEF ≥50%), n (%)	33 (42.9)	246 (50.3)	154 (48.4)	0.46
HFmrEF (40% ≤ LVEF < 50%), n (%)	11 (14.3)	79 (16.2)	56 (17.6)	0.74
HFrEF (LVEF <40%), n (%)	33 (42.9)	164 (33.5)	108 (34.0)	0.27

Values are presented as number (%), mean±standard deviation, or median [25th, 75th percentiles]. ACEI, angiotensin-converting enzyme inhibitor; ARB, angiotensin-II receptor blocker; BMI, body mass index; BNP, B-type natriuretic peptide; COPD, chronic obstructive pulmonary disease; CV, cardiovascular; DOAC, direct oral anticoagulant; DPP, dipeptidyl peptidase; eGFR, estimated glomerular filtration rate; HFpEF, heart failure preserved ejection fraction; HFmrEF, heart failure with mid-range ejection fraction; HFrEF, heart failure reduced ejection fraction; LDL-C, low-density lipoprotein cholesterol; LVEF, left ventricular ejection fraction; MRA, mineralocorticoid receptor antagonist; NYHA, New York Heart Association; NSAID, nonsteroidal anti-inflammatory drug; SGLT, sodium glucose cotransporter.

Figure 2.

Distribution of (A) the total number of medications, (B) CV medications, and (C) non-CV medications. CV, cardiovascular; IQR, interquartile range; SD, standard deviation.

Figure 3.

Median number of class-specific medications stratified as CV and non-CV. CV, cardiovascular.

During the median follow-up period of 730 days (interquartile range, 228–730 days), all-cause death was observed in 232 patients (incidence rate: 15.6 per 100 patient-years). In the Kaplan-Meier analysis, patients in the hyperpolypharmacy group showed worse outcomes than patients in the other 2 groups did (log-rank P=0.002; Figure 4A). Compared with patients taking ≤8 (median) medications, those taking >8 (median) medications also demonstrated worse outcomes (log-rank P=0.029; Figure 4B). Stratified by the medication class (CV medications vs. non-CV medications and HF medications vs. non-HF medications), patients taking >5 (median) CV medications were not observed to have an increased mortality rate (log-rank P=0.64); however, compared with patients taking ≤3 (median) non-CV medications, those taking >3 (median) non-CV medications showed worse prognoses (log-rank P<0.001). We also observed that patients taking >3 (median) HF medications did not show an increased risk of death (log-rank P=0.20); however, compared with patients taking ≤5 non-HF medications, those taking >5 (median) non-HF medications showed worse prognoses (log-rank P=0.10; Figure 5). Table 2 presents the results of the univariable and multivariable adjusted Cox proportional hazards analyses for all-cause death, categorized by the polypharmacy status. In the crude model, compared with polypharmacy, hyperpolypharmacy was associated with a higher risk of all-cause death (HR, 1.57; 95% CI, 1.20–2.05; P<0.001). After multivariable adjustment for the 21 covariates, hyperpolypharmacy remained significantly associated with an increased risk of all-cause death (HR, 1.48; 95% CI, 1.11–1.99; P=0.008). Furthermore, the total number of medications also remained significantly associated with an increased incidence of all-cause death (HR per additional increase in the number of medications, 1.05; 95% CI, 1.01–1.10; P=0.027). Although an additional increase in the number of non-CV medications was significantly associated with all-cause death (HR per additional increase in the number of non-CV medications, 1.07; 95% CI, 1.02–1.13; P=0.01), an additional increase in the number of CV medications was not (HR per additional increase in the number of CV medications, 1.01; 95% CI, 0.92–1.10; P=0.95). As for HF medications, similar results were found (HR per additional increase in the number of non-HF medications, 1.06; 95% CI, 1.00–1.11; P=0.034; HR per additional increase in the number of HF medications, 1.04; 95% CI, 0.93–1.17; P=0.47; Table 3). We assessed the associations of hyperpolypharmacy and nonhyperpolypharmacy with all-cause death in terms of the baseline characteristics. The results were fundamentally the same (Supplementary Tables 3,4). Subgroup analyses revealed no significant interactions of hyperpolypharmacy and age, sex, hypertension, diabetes mellitus, atrial fibrillation, or LVEF with all-cause death (all P for interaction >0.05). In contrast, we observed significant interactions of hyperpolypharmacy and the NYHA functional class at discharge (NYHA I–II vs. III–IV) and etiology (ischemic cardiopathy vs. nonischemic) with all-cause death (P=0.007 and 0.042, respectively), suggesting modification of the effect of hyperpolypharmacy on all-cause death by a severe HF status or ischemic etiology (Supplementary Table 5, Supplementary Figure). The baseline clinical characteristics according to the NYHA functional class and etiology are listed in Supplementary Tables 6 and 7, respectively. The proportions of patients with severe HF (NYHA classes III–IV) and those with ischemic cardiomyopathy was higher in the HF group with a reduced EF, suggesting that the medications prescribed in accordance with practice guidelines led to an increase in the number of medications consumed; however, there was no definitive association of hyperpolypharmacy with death in these subpopulations.

Figure 4.

Kaplan-Meier curves in 3 groups categorized by polypharmacy status (A) and by the median total number of medications (>8 medications vs. ≤8 medications) (B).

Figure 5.

Kaplan-Meier curves in 2 groups by (A) the median total number of CV medications (>5 medications vs. ≤5 medications), total number of non-CV medications (>3 medications vs. ≤3 medications) (B), total number of HF medications (>3 medications vs. ≤3 medications) (C), and total number of non-HF medications (>5 medications vs. ≤5 medications) (D). CV, cardiovascular; HF, heart failure.

Table 2. Cox Proportional Hazards Analyses of Clinical Outcomes by Polypharmacy Category

Polypharmacy category	Overall (n=884) (median, [25th, 75th percentiles])	Nonpolypharmacy (<5 medications) (n=77) HR (95% CI) P value (vs. Poly-)	Polypharmacy (5–9 medications) (n=489) Reference	Hyperpolypharmacy (≥0 medications) (n=318) HR (95% CI) P value (vs. Poly-)	Medication, per 1 increase HR (95% CI)
All-cause death (Follow-up days)	232 events (730 [603–730])	17 events	110 events	105 events
Event rate (per 100 patient-years)	15.6 (13.7–17.8)	12.4 (7.7–19.9)	13.2 (10.9–15.9)	20.5 (16.9–24.8)
Unadjusted (n=884)		0.97 (0.58–1.61)	Reference	1.57 (1.20–2.05)	1.07 (1.03–1.11)
		P=0.89		P<0.001	P<0.001
Adjusted (n=884)		1.32 (0.77–2.26)	Reference	1.48 (1.11–1.99)	1.05 (1.01–1.10)
		P=0.31		P=0.008	P=0.027

Multivariable model was adjusted for age, sex, LVEF, NYHA functional class at discharge, prior hospitalization for heart failure, hypertension, dyslipidemia, diabetes mellitus, hyperuricemia, atrial fibrillation, history of coronary artery disease, history of cerebrovascular disease, history of peripheral artery disease, obesity (BMI ≥30 kg/m²), chronic obstructive pulmonary disease, dementia, malignant tumor, hemoglobin, albumin, eGFR, and BNP. CI, confidence interval; HR, hazard ratio. Other abbreviations as in Table 1.

Table 3. Cox Proportional Hazards Analyses of Clinical Outcome (All-Cause Death) by CV Medications, Non-CV Medications, HF Medications, and Non-HF Medications

Medication class	Medication, per 1 increase HR (95% CI), P value
CV medications
Unadjusted (n=884)	0.98 (0.91–1.05), P=0.53
Adjusted (n=884)	1.01 (0.92–1.10), P=0.95
Non-CV medications
Unadjusted (n=884)	1.12 (1.07–1.18), P<0.001
Adjusted (n=884)	1.07 (1.02–1.13), P=0.01
HF medications
Unadjusted (n=884)	0.95 (0.86–1.06), P=0.36
Adjusted (n=884)	1.04 (0.93–1.17), P=0.47
Non-HF medications
Unadjusted (n=884)	1.10 (1.05–1.14), P<0.001
Adjusted (n=884)	1.06 (1.00–1.11), P=0.034

HF medications: ACEIs, ARBs, mineralocorticoid receptor antagonists, β-blockers, diuretics (loop, thiazide, and tolvaptan), digoxin, amiodarone, nitrates, nicorandil, and sodium glucose cotransporter-2 inhibitors. Multivariable model was adjusted for age, sex, LVEF, NYHA functional class at discharge, prior hospitalization for heart failure, hypertension, dyslipidemia, diabetes mellitus, hyperuricemia, atrial fibrillation, history of coronary artery disease, history of cerebrovascular disease, history of peripheral artery disease, obesity (BMI ≥30 kg/m²), chronic obstructive pulmonary disease, dementia, malignant tumor, hemoglobin, albumin, eGFR, and BNP. Abbreviations as in Tables 1,2.

Discussion

In this post-hoc analysis of the CURE-HF database, >90% and 36% of the patients after ADHF met the commonly accepted definitions of polypharmacy (≥5 medications) and hyperpolypharmacy (≥10 medications), respectively. We demonstrated that hyperpolypharmacy reflected multiple comorbidities and was associated with a higher risk of death in patients after ADHF. We found that the number of non-CV medications, but not the number of CV medications, was significantly associated with death.

The prevalence of polypharmacy and hyperpolypharmacy was reported to be 93% and 54%, respectively, in the TOPCAT trial,¹³ and 95% and 55%, respectively in patients hospitalized for HF in a USA-based study.⁵ The prevalence of hyperpolypharmacy in those studies was substantially higher than in our study of patients after ADHF. Although we did not include over-the-counter medications or vitamins, these medications were assessed in the TOPCAT trial and may have partially contributed towards the observed differences.^5,13 Our study found that patients with polypharmacy and hyperpolypharmacy had more comorbidities, consistent with the findings of other studies of HF patients.^{5,13–15,23,24} Polypharmacy in patients with HF can occur for several reasons, including guideline-directed medical therapies and treatments of multiple comorbidities linked to aging. In this study, classification of medications showed that as the number of total medications increased, the number of non-CV medications increased significantly, more than that of CV medications (Figure 3). The distribution trends for the total and non-CV medications were similar. Unlu et al reported that the number of non-CV medications increased more than the number of CV medications in older adult patients hospitalized for HF,⁵ which concords with our results.

Prior studies of HF patients have explored the associations between hyperpolypharmacy and clinical outcomes. In the KCHF registry, hyperpolypharmacy was associated with a higher risk of death within a 1-year period, consistent with our results.¹⁴ In the TOPCAT trial, post-hoc analysis restricted to the Americas revealed that hyperpolypharmacy was not significantly associated with death, but was significantly associated with hospitalization for any reason and any serious adverse events during the median 2.9-year follow-up.¹³ Although the reasons for these discrepant findings are unclear, differences in patients’ ages and mortality rates among the studies is a likely cause. The median ages of the patients in the CURE-HF registry, KCHF registry, and TOPCAT trial were 81, 77, and 72 years, respectively, and the mortality rates were 15.6, 10.5, and 7.1 (per 100 patient-years), respectively. The median ages in the CURE-HF and KCHF registries were 5–9 years higher than in the TOPCAT trial, which may, in part, account for the difference. Additionally, variability in the mortality rate might have contributed to the observed differences in the results from the 3 studies. Further, older patients with HF more often died, giving a higher mortality rate in our study than in the TOPCAT trial. Hence, hyperpolypharmacy remained significantly associated with an increased risk of death after adjustment for clinical risk factors in both our study and the KCHF registry. Additionally, we assessed the associations between the classification of medications (CV medications vs. non-CV medications and HF medications vs. non-HF medications) and death in patients with HF. We clarified that an increase in the CV or HF medications was not significantly associated with death, suggesting that prescription of these medications under guideline-directed medical treatment for HF is relevant and safe in clinical practice. Indeed, we found that the proportions of patients with severe HF (NYHA classes III–IV) and those with ischemic cardiomyopathy were higher in the HF with reduced EF group, which suggests the importance of guideline-directed medical therapies, irrespective of an absolute increase in the total medications or HF medications. A recent study demonstrated that an intensive treatment strategy based on guideline-directed medical therapies and close follow-up after ADHF was associated with a reduction in CV outcomes at the 180-day follow-up.²⁵ Therefore, our exploratory findings suggest that recognizing the risk factors and establishing optimal guideline-directed medical therapies could reduce the HF burden or at least prevent an increase in future risks for CV outcomes despite an increase in the CV or HF medications.

On the other hand, we found that an increase in the non-CV medications remained associated with death after adjustment for clinical risk factors (including comorbidities). The older participants were prescribed fewer CV medications and more non-CV medications.⁵ Patients with hyperpolypharmacy are more likely to have a variety of comorbidities, which often reflects an advanced phase of systemic illness and may have contributed to our findings.

Poor medication adherence is also associated with adverse outcomes in patients with polypharmacy. Intervention by healthcare providers at hospital discharge contributed to an improvement in medication compliance.²⁶ Thus, multidisciplinary interventions at discharge are an important first step to optimizing prescribing practices. Our study clarified the effect of hyperpolypharmacy on post-ADHF prognosis by drug classification (CV medications vs. non-CV medications and HF medications vs. non-HF medications), suggesting that evaluation of the ADHF patients’ polypharmacy status may be useful for identifying that at an elevated risk of death. Our findings provide a basis for future prospective randomized studies to evaluate the role of multidisciplinary interventions in prescribing medications at discharge and subsequent reviewing of clinical outcomes in patients after ADHF.

Study Limitations

This study should be considered in the context of its limitations. First, it was not a prespecified analysis, and unmeasured confounding factors might have affected the results, regardless of the multivariable adjustment. Second, the polypharmacy status was only evaluated at discharge, and we did not assess changes in the total number of medications or medication patterns during the follow-up period. We did not have data on the dosage of the medications at baseline. Third, the registration period was relatively long and standard treatments changed slightly over time. In recent years, SGLT-2 inhibitors have been used as a standard medication for HF, and in this study we defined them as CV medications, even though they were used as hypoglycemic drugs during the registration period and were generally defined as non-CV medications in previous studies. Fourth, we were unable to assess the adverse effects of polypharmacy, such as adverse drug interactions and non-adherence, because this study was not designed to identify causality between hyperpolypharmacy and adverse events in patients after ADHF; thus, our exploratory findings should be considered as hypothesis-generating. Fifth, potentially inappropriate medications (PIMs) are known to be associated with adverse drug events in patients taking ≥5 medications.²⁴ PIMs have also been reported to be more common in older adult patients hospitalized for HF.²⁷ In our registry, we assessed PIMs based on the Beers 2019 criteria,²⁸ and 5 medications (digitalis, amiodarone, proton pump inhibitors, nonsteroidal anti-inflammatory drugs [NSAIDs], and pioglitazone) were included as PIMs, although we were not able to determine whether digitalis and amiodarone were used as first-line agents for atrial fibrillation (Supplementary Table 7). After adjusting for demographics and comorbidities, we observed that each PIM and the total number of PIMs did not cause worse outcomes. Our findings could be primarily driven by the crude number of medications rather than by certain drug patterns; however, we lacked precise breakdown data on medications such as antiplatelet agents, antipsychotic drugs, and sleeping pills, making a detailed assessment of PIMs difficult. Finally, we did not assess whether our results were relevant irrespective of the HF subtype (HF with reduced EF vs. HF with preserved EF) due to the sample size. Guideline-directed medical therapies for HF with reduced EF are associated with improvement in CV outcomes, and there are different medical strategies based on the LVEF;⁴ however, there was no significant interaction between hyperpolypharmacy and the HF subtype (LVEF <40% vs. LVEF ≥40%) in the present study.

In conclusion, hyperpolypharmacy was common in patients after ADHF in the CURE-HF registry, and was associated with an elevated risk of all-cause death. The number of non-CV medications was strongly associated with an increased risk of death, but the number of CV medications was not. Our post-hoc exploratory findings suggest the importance of prescribing drugs, including non-CV medications, with regular reviews to optimize the guideline recommendations for treatment of patients after ADHF.

Acknowledgments

The authors thank all the participants (patients, caregivers, and staff) in the CURE-HF Registry and Minako Aono for their invaluable contributions.

This study was supported by a grant (no. 18059186) from the Japanese Agency for Medical Research and Development (Drs. H. Motoki, N. Ozasa, Takao Kato, and K. Kuwahara). The funder had no role in the study design and collection, analysis, or interpretation of the data; writing of the manuscript; or the decision to submit the article for publication.

Disclosures

Dr. K. Kuwahara is a member of Circulation Journal’s Editorial Team. The other authors declare they have no conflicts of interest.

IRB Information

The present study was approved by the Institutional Review Board of the Shinshu University School of Medicine (approval no. 4237).

Data Availability Statement

The de-identified participant data will not be shared.

Supplementary Files

Please find supplementary file(s);

https://doi.org/10.1253/circj.CJ-22-0712

References

• 1.   Shah SJ, Kitzman DW, Borlaug BA, van Heerebeek L, Zile MR, Kass DA, et al. Phenotype-specific treatment of heart failure with preserved ejection fraction: A multiorgan roadmap. Circulation 2016; 134: 73–90.
• 2.   Shah AM, Claggett B, Loehr LR, Chang PP, Matsushita K, Kitzman D, et al. Heart failure stages among older adults in the community: The atherosclerosis risk in communities study. Circulation 2017; 135: 242–240.
• 3.   Shimokawa H, Miura M, Nochioka K, Sakata Y. Heart failure as a general pandemic in Asia. Eur J Heart Fail 2015; 17: 884–892.
• 4.   Tsutsui H, Isobe M, Ito H, Ito H, Okumura K, Ono M, et al. JCS 2017/JHFS 2017 guideline on diagnosis and treatment of acute and chronic heart failure-digest version. Circ J 2019; 83: 2084–2184.
• 5.   Unlu O, Levitan EB, Reshetnyak E, Kneifati-Hayek J, Diaz I, Archambault A, et al. Polypharmacy in older adults hospitalized for heart failure. Circ Heart Fail 2020; 13: e006977.
• 6.   McDonagh TA, Metra M, Adamo M, Gardner RS, Baumbach A, Böhm M, et al. 2021 ESC guidelines for the diagnosis and treatment of acute and chronic heart failure. Eur Heart J 2021; 4: 3599–3726.
• 7.   Heidenreich PA, Bozkurt B, Aguilar D, Allen LA, Byun JJ, Colvin MM, et al. 2022 AHA/ACC/HFSA Guideline for the management of heart failure: A report of the American College of Cardiology/American Heart Association joint committee on clinical practice guidelines. Circulation 2022; 145: e895–e1032.
• 8.   Runganga M, Peel NM, Hubbard RE. Multiple medication use in older patients in post-acute transitional care: A prospective cohort study. Clin Interv Aging 2014; 9: 1453–1462.
• 9.   Nishtala PS, Salahudeen MS. Temporal trends in polypharmacy and hyperpolypharmacy in older New Zealanders over a 9-year period: 2005–2013. Gerontology 2015; 61: 195–202.
• 10.   Masnoon N, Shakib S, Kalisch-Ellett L, Caughey GE. What is polypharmacy? A systematic review of definitions. BMC Geriatr 2017; 17: 230.
• 11.   Marcum ZA, Amuan ME, Hanlon JT, Aspinall SL, Handler SM, Ruby CM, et al. Prevalence of unplanned hospitalizations caused by adverse drug reactions in older veterans. J Am Geriatr Soc 2012; 60: 34–41.
• 12.   Gómez C, Vega-Quiroga S, Bermejo-Pareja F, Medrano MJ, Louis ED, Benito-León J. Polypharmacy in the elderly: A marker of increased risk of mortality in a population-based prospective study (NEDICES). Gerontology 2015; 61: 301–309.
• 13.   Minamisawa M, Claggett B, Suzuki K, Hegde SM, Shah AM, Desai AS, et al. Association of hyperpolypharmacy with clinical outcomes in heart failure with preserved ejection fraction. Circ Heart Fail 2021; 14: e008293.
• 14.   Ozasa N, Kato T, Morimoto T, Yaku H, Yamamoto E, Inuzuka Y, et al. Polypharmacy and clinical outcomes in hospitalized patients with acute decompensated heart failure. J Cardiovasc Nurs 2023; 38: 33–43.
• 15.   Okano T, Motoki H, Minamisawa M, Kimura K, Kanai M, Yoshie K, et al. Cardio-renal and cardio-hepatic interactions predict cardiovascular events in elderly patients with heart failure. PLos One 2020; 15: e0241003.
• 16.   Suzuki S, Motoki H, Kanzaki Y, Maruyama T, Hashizume N, Kozuka A, et al. Prognostic impact of mineralocorticoid receptor antagonist in patients with heart failure with preserved ejection fraction. ESC Heart Fail 2020; 7: 2752–2761.
• 17.   Sakai T, Motoki H, Suzuki S, Fuchida A, Takeuchi T, Otagiri K, et al. Gender difference in heart failure with preserved ejection fraction: Clinical profiles, examinations, and prognosis. Heart Vessels 2022; 37: 1710–1718.
• 18.   McMurray JJV, Solomon SD, Inzucchi SE, Kober L, Kosiborod MD, Martinez FA, et al. Dapagliflozin in patients with heart failure and reduced ejection fraction. N Engl J Med 2019; 381: 1995–2008.
• 19.   Packer M, Anker SD, Butler J, Filippatos G, Ferreira JP, Pocock SJ, et al. Effect of empagliflozin on the clinical stability of patients with heart failure and a reduced ejection fraction: The EMPEROR-reduced trial. Circulation 2021; 143: 326–336.
• 20.   Packer M, Butler J, Zannad F, Filippatps G, Pocock SJ, Carson P, et al. Effect of empagliflozin on worsening heart failure events in patients with heart failure and preserved ejection fraction: EMPEROR-preserved trial. Circulation 2021; 144: 1284–1294.
• 21.   Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: Development and validation. J Chronic Dis 1987; 40: 373–383.
• 22.   Elixhauser A, Steiner C, Harris DR, Coffey RM. Comorbidity measures for use with administrative data. Med Care 1998; 36: 8–27.
• 23.   Robert LP, Bryant CL, Cheng D, Dow TJ, Ky B, Stein CM, et al. Drug that may cause or exacerbate heart failure. Circulation 2016; 134: e32–e69.
• 24.   Tommelein E, Mehuys E, Petrovic M, Somers A, Colin P, Boussery K. Potentially inappropriate prescribing in community dwelling older people across Europe: A systematic literature review. Eur J Clin Pharmacol 2015; 71: 1415–1427.
• 25.   Mebazaa A, Davison B, Chioncel O, Cohen-Solal A, Diaz R, Filippatos G, et al. Safety, tolerability and efficacy of up-titration of guideline-directed medical therapies for acute heart failure (STRONG-HF): A multinational, open-label, randomised, trial. Lancet 2022; 400: 1938–1952.
• 26.   Mortelmans L, Baetselier ED, Goossens E, Dilles T. What happens after hospital discharge? Deficiencies in medication management encountered by geriatric patients with polypharmacy. Int J Environ Res Public Health 2021; 18: 7031.
• 27.   Jaber D, Vargas F, Nguyen L, Ringel J, Zarziela K, Musse M, et al. Prescriptions for potentially inappropriate medications from the Beers criteria among older adults hospitalized for heart failure. J Card Fail 2022; 8: 906–915.
• 28.   2019 American Geriatrics Society Beers Criteria® Update Expert Panel Fick DM, SemLa TP, Steinman M, Beizer J, Brandt N, et al. American Geriatrics Society 2019 Updated AGS Beers Criteria® for potentially inappropriate medication use in older adults. J Am Geriatr Soc 2019; 67: 674–694.